Grade boundaries are the tool used by exam boards to award grades to students. The standard associated with each grade at GCSE, AS, A Level is set and regulated across subjects; the aim is that a 5 in Maths GCSE is worth the same as a 5 in Latin or Classical Civilisation for example. The boundaries therefore have to be flexible to make sure this standard is protected and that grades awarded to students are meaningful and valid.
New specification, new papers, new boundaries.
Boundaries are set after a paper is sat and all marking is complete so that they reflect actual student performance. Papers are not designed to produce certain grade boundaries, the person designing or setting the paper may have a notion of where they might expect the boundary to be, but that’s as far as it goes. This is also why before a paper is sat, the exam board can’t give the grade boundary, as they don’t know how students will perform on the paper. This might seem annoying but it’s a good thing; if the boundaries were set in advance it would affect the reliability of the results as they wouldn’t reflect the reality of how your students responded to the paper.
Those handling awarding do so by looking at the standard of work produced and aligning it with the known standard associated with a particular grade. It’s worth noting that this is done at paper level, not question by question. To put it simply, work that would usually be of A grade standard should fall above the A grade boundary, work of an E grade standard should fall above the E grade boundary but below the D grade one. A mix of judgements such as these and statistical analysis – including checking against the cohort’s predicted performance based on prior attainment data – are used to ascertain where to put the boundaries.
Grade boundaries are dependent on the specific subject and paper. Therefore comparing boundaries from the old specification or boundaries from other subjects is of no use; they will have no impact on the boundaries of the qualification you’re looking at and will not create meaningful predictions. Historic data for a specification can be used as a predictor (cohorts don’t differ wildly year on year and so neither should boundaries) but between specifications, one boundary can’t be used to predict another. This of course means that predicting grades for a brand new specification is incredibly difficult.
So that’s the unhelpful part. Now the (hopefully) helpful advice.
Use your own professional judgement, much as the principal examiners will have to. Take the papers you have marked and put them in rank order, worst to best. Think about the quality of work that you would historically have considered to be ‘C’ grade, and look at where in your marks that seems to lie. You may find looking at the level descriptors in the mark scheme helpful (‘sound’ or ‘reasonable’ are often words associated with C-ish work, ‘very good’, ‘detailed’ or ‘well-chosen evidence’ are all terms more associated with A grade work). Then do the same for A and E. Place these three boundaries at points that correspond to the student work in front of you and the language of the level descriptors. Then spread the other grades evenly between them.
To check your thoughts, look at the distribution of your students. If you have a mixed ability class you’d expect a pretty normal distribution: most students getting the middling grades, fewer at either end. If you have a skewed cohort, you’d expect the grade distribution to skew accordingly.
This is much more work than a simple solution, and it’s not a precise science, but it is a good way to create meaningful grade predictions for students.
At A Level the standard hasn't changed and the intention is not that the qualifications get harder. Work which previously was considered to be A grade standard should still get an A and so on. It’s this understanding which will inform the real awarding process. So if you are setting your boundaries as described above you can use your judgement based on the quality of work you’ve seen in previous years. What you recognise as C grade work should still be getting a C.
For GCSE it is a little trickier as the standard has changed and there are brand new grades to contend with. Ofqual published a handy chart showing how the old and new grades align. This can help you when you are doing your students’ predictions. To set the 4 and 7 boundaries (C and A) use your judgement to say what standard of work would constitute a low C or A historically and set the 4 and 7 boundaries to reflect this. Have a look at OCR’s website for e for the live series and a good explanation of all the issues.
No. Boundaries aren't supposed to be 'aligned' across subjects – although the standards associated with different grades are – so difference between subjects isn't a concern. There isn't a 'boundary blueprint'. It is more concerning if boundaries vary across years in one qualification, that means the paper difficulty is fluctuating.
Low boundaries don’t mean more high grades, high boundaries don’t mean it is harder to get good grades. A high grade boundary likely reflects a paper that students found relatively straightforward and achieved a lot of marks on. Lower boundaries are probably due to a more challenging paper. The boundaries are, however, tied to the overall standard; so if a piece of work deserved an A grade it should be above the A grade threshold wherever that is located.
All boards and subjects are regulated by Ofqual in terms of their outcomes. If one board or subject has a spike in higher grades which is out of line with expected student performance they will investigate and possibly alter the grade boundaries to redress the balance. The aim is that all qualifications and subjects should be of equal difficulty and the standard should be the same in all of them.
Grade boundaries and outcomes are designed to be fair. No student will be disadvantaged or advantaged, the system is designed to try and ensure they would have got the same grade for similar work regardless of what specification they are using.
No. Boundaries take into account the nature of the cohort. This can be worked out using prior attainment data. It can be relatively easily ascertained when a cohort sitting an exam is skewed one way or the other and steps can be taken to address this.
It is important to remember that grade boundaries are set based on the standard of student work, not how many students are ‘allowed’ to get a particular grade. If a cohort is strong, then more students get given high grades. If it is weak, fewer do. There is no cap on how many students can get an A or a B for example, if examiners see a lot of A grade work then lots of students are getting As that year!
Calculation of the Grade 9 boundary at GCSE is also tied to the nature of a subject’s cohort, therefore it is not harder to get a 9 in Classical subjects.
